happy CGN -- beating happy eyeballs and trending toward E2E success

[Cameron Byrne]

On Jun 12, 2012 4:38 PM, "Andrew Dickinson" <whydna at whydna.net> wrote:
>
> Is there any reason to have a delay at all (at least for TCP)?
>
> Would it be unreasonable for a client implementation to simultaneously
> send a SYN on both v4 and v6?  It would then issue a RST for the
> SYN-ACK that arrives last.  I recognize that this may need to be done
> at the application layer which probably makes it a no-go of a
> generalized solution, but I suspect it could be incorporated into an
> OS's TCP stack.
>
> -A
>

As noted at the top of the thread, network access providers and content
providers do not want a straight race, they want end to end v6 and they
want to avoid cgn.

Apple's happy eyeballs implementation leads to too much ipv4 via nat444 and
not enough ipv6 e2e.

Cgn does not introduce delay in most networks.

Thus, we need to explore options for meeting the content and network
providers real business needs, not some theoretical users experience
benefits afforded from a straight race where ipv4 is 5ms faster (cannot be
perceived)

Fyi, this is really only a problem for apple products. The rest of the
world prefers ipv6.

CB

> On Tue, Jun 12, 2012 at 3:50 PM, Daniel Roesen <dr at cluenet.de> wrote:
> > On Tue, Jun 12, 2012 at 05:21:05PM -0400, Erik Nygren wrote:
> >> A "good" Happy Eyeballs implementation should compare IPv4 vs IPv6
> >> TCP (HTTP) performance and availability with some preference towards
> >> IPv6 (whether it is 30ms or 300ms).
> >
> > Unfortunately, Apple thinks different:
> >
> > http://www.ietf.org/mail-archive/web/v6ops/current/msg10080.html
> >
> > Best regards,
> > Daniel
> >
> > --
> > CLUE-RIPE -- Jabber: dr at cluenet.de -- dr at IRCnet -- PGP: 0xA85C8AA0
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.cluenet.de/pipermail/ipv6-ops/attachments/20120612/0d2abb63/attachment.html